KR102454701B1 - Device manufacturing method - Google Patents

Abstract

A device manufacturing method comprising: acquiring a time series of measurement data of a plurality of substrates on which an exposure step and a process step have been performed; obtaining a time series of state data related to a condition prevailing when a process step has been performed on at least a portion of the plurality of substrates; applying a filter to the measurement data time series and the state data time series to obtain filtered data; and using the filtered data to determine a correction to be applied to an exposure step performed on a subsequent substrate.

Description

Device manufacturing method

CROSS-REFERENCE TO RELATED APPLICATIONS

This application claims priority to EP Application No. 17175967.3, filed on June 14, 2017 and incorporated herein by reference in its entirety.

BACKGROUND OF THE INVENTION Field of the Invention The present invention relates to device fabrication using a lithographic apparatus and a processing apparatus.

A lithographic apparatus is an apparatus that imparts a desired pattern onto a substrate, typically on a target area of the substrate. The lithographic apparatus may be used, for example, in the manufacture of integrated circuits (ICs). In that case, a patterning device, also called a mask or reticle, may be used to create a circuit pattern to be formed on individual layers of the integrated circuit. Such a pattern may be transferred onto a target portion (eg, including a portion of a die, one die, or several dies) on a substrate (eg, a silicon wafer). Transfer of the pattern is typically accomplished via imaging onto a layer of radiation-sensitive material (resist) provided on a substrate. In general, a single substrate will contain a network of adjacent target portions that are successively patterned. The patterned radiation-sensitive layer is then developed, and a process device such as an etcher is used to immobilize the pattern into the substrate.

In order to manufacture an electronic device, it is necessary to repeat the exposure and fixing steps, for example up to 30 times, to create different layers of the device. Each layer is applied at a time in a batch, also known as a lot of substrates. To improve yield, i.e., proportion of functional devices, by adjusting exposure of subsequent substrates in the same batch or subsequent batches to which the same process is applied using measurements performed on the substrate, for example overlay, focus or It is known to reduce errors in CD. This process is known as automated process control. When measurements of multiple substrates are possible, a weighted moving average is often used as a measurement to process control.

However, known APC methods still leave "fingerprints", i.e. variations in parameters of focus, overlay or CD across the substrate, and there is therefore a need to improve automated process control methods.

SUMMARY OF THE INVENTION It is an object of the present invention to provide an improved automated process control method for use in a lithographic manufacturing process.

In a first aspect, the present invention provides a device manufacturing method comprising:

acquiring a time series of measurement data of a plurality of substrates on which the exposure step and the process step have been performed;

obtaining a time series of state data related to a condition prevailing when a process step has been performed on at least a portion of the plurality of substrates;

applying a filter to the measurement data time series and the state data time series to obtain filtered data; and

and using the filtered data to determine a correction to be applied to an exposure step performed on a subsequent substrate.

In a second aspect, the present invention provides a device manufacturing method comprising:

acquiring a time series of measurement data of a plurality of substrates on which the exposure step and the process step have been performed;

applying a filter to the measurement data time series to obtain filtered data;

using the filtered data to determine a correction to be applied to an exposure step performed on a subsequent substrate;

applying an additional filter to the measurement data time series to obtain additional filtered data; and

and using the further filtered data to determine further corrections to be applied to the exposure step performed on the subsequent substrate.

In a third aspect, the present invention provides a device manufacturing method comprising:

acquiring a time series of measurement data of a plurality of substrates on which the exposure step and the process step have been performed;

obtaining a time series of state data related to a condition prevailing when a process step has been performed on at least a portion of the plurality of substrates;

converting the measurement data time series and the state data time series into frequency space data;

determining a filter to be applied to at least one of the measurement data time series and the state data time series based on the frequency spatial data to obtain filtered data;

applying the filter to at least one of the measurement data time series and the state data time series to obtain filtered data; and

and using the filtered data to determine a correction to be applied to an exposure step performed on a subsequent substrate.

In a fourth aspect, the present invention provides a device manufacturing method comprising:

acquiring a time series of measurement data of a plurality of substrates on which the exposure step and the process step have been performed;

obtaining a time series of state data related to a condition prevailing when a process step has been performed on at least a portion of the plurality of substrates;

converting the measurement data time series and the state data time series into frequency space data; and

and determining, based on the frequency spatial data, a sampling scheme to be applied to a subsequent substrate to generate a time series of measurement data.

Now, embodiments of the present invention will be described by way of illustration only with reference to the accompanying drawings:
1 shows a lithographic apparatus together with another apparatus for manufacturing a production facility for a semiconductor device;
2 shows a conventional automated process control method;
3 shows the principle of operation of an automated process control method according to an embodiment of the present invention;
4 shows a specific application of an automated process control method according to an embodiment of the present invention;
5 shows a specific application of an automated process control method according to an embodiment of the present invention;
6 shows a specific application of an automated process control method according to an embodiment of the present invention;
7 shows a specific application of an automated process control method according to an embodiment of the present invention;
8 shows a process for establishing an automated process control method according to an embodiment of the present invention; and
9 shows a simulated example of the effect of a process parameter on a substrate fingerprint.

Before describing embodiments of the present invention in detail, it is advantageous to present an exemplary environment in which embodiments of the present invention may be implemented.

1 shows a typical layout of a semiconductor production facility. The lithographic apparatus 100 applies a desired pattern onto a substrate. Lithographic apparatus is used, for example, in the manufacture of integrated circuits (ICs). In that case, the patterning device MA, also referred to as a mask or reticle, contains a circuit pattern of features (often referred to as "product features") to be formed on individual layers of the IC. Through exposure 104 of the patterning device onto a layer of radiation-sensitive material (resist) provided on the substrate, such a pattern is formed in a target portion (eg, a die) on a substrate 'W' (eg, a silicon wafer). part of the dies, including one or several dies). In general, a single substrate will contain a network of adjacent target portions that are successively patterned.

A known lithographic apparatus illuminates each target portion by illuminating the patterning device while positioning the target portion of the substrate synchronously with the image position of the patterning device. The irradiated target portion of the substrate is referred to as the "exposure field", or simply "field." The layout of a field on a substrate is typically a network of adjacent rectangles aligned according to a Cartesian two-dimensional coordinate system (eg aligned along the X and Y-axes, both axes being orthogonal to each other).

A requirement for a lithographic apparatus is an accurate and accurate reproduction of a desired pattern on a substrate. The location and dimensions of the applied product features need to fall within certain tolerances. Positional errors can be caused by overlay errors (often called "overlays"). Overlay is an error in positioning a first product feature in a first layer relative to a second product feature in a second layer. The lithographic apparatus minimizes overlay errors by accurately aligning each wafer to a reference prior to patterning. This is done by measuring the position of an alignment mark applied to the substrate. Based on the alignment measurements, the substrate position is controlled during the patterning process to prevent overlay errors from occurring.

Errors in the critical dimension (CD) of a product feature may occur if the applied dose associated with exposure 104 is not within specifications. For this reason, the lithographic apparatus 100 must be able to accurately control the dose of radiation applied to the substrate. CD errors can also occur if the substrate is not positioned correctly with respect to the focal plane associated with the pattern image. The focal position error is generally associated with the non-planarity of the substrate surface. The lithographic apparatus minimizes this focal position error by measuring the substrate surface topography using a level sensor prior to patterning. Substrate height correction is applied during subsequent patterning to ensure accurate imaging (focusing) of the patterning device onto the substrate.

The patterned substrate is measured by metrology apparatus 140 to verify overlay and CD errors associated with the lithography process. A common example of a metrology device is a scatterometer. Traditionally, scatterometers measure the characteristics of a dedicated metrology target. These metrology targets are representative of product features except that their dimensions are typically larger to allow for accurate measurements. Scatterometers measure overlay by detecting asymmetry in the diffraction pattern associated with the overlay metrology target. The critical dimension is determined by analysis of the diffraction pattern associated with the CD metrology target. Another example of a metrology tool is an electron beam (e-beam) based inspection tool, such as a scanning electron microscope (SEM).

Within a semiconductor production facility, the lithographic apparatus 100 and metrology apparatus 140 form part of a "lithocell" or "litho cluster". The litho cluster includes a coating device 108 for applying a photosensitive resist to a substrate W, a baking device 110, a developing device 112 for developing the exposed pattern into a physical resist pattern, an etching station 122, It further includes an apparatus 124 for performing a post-etch annealing step and possibly other processing apparatus 126 and the like. The metrology apparatus is configured to inspect the substrate after development 112 or after further processing (eg, etching). The various apparatuses within the lithocell are controlled by a supervisory control system (SCS), which issues a control signal 166 for controlling the lithographic apparatus via a lithographic apparatus control unit LACU 106 to perform a recipe R. . SCS allows other devices to operate while providing maximum throughput and product yield. An important control mechanism is feedback 146 (via SCS) of metrology apparatus 140 to various apparatuses, particularly lithographic apparatus 100 . Based on the characteristics of the metrology feedback, corrective actions are determined to improve the processing quality of subsequent substrates.

Conventionally, the performance of a lithographic apparatus is controlled and corrected by methods such as for example automated process control (APC) described in US201208127A1. Automated process control techniques use measurement of metrology targets applied to a substrate. A Manufacturing Execution System (MES) schedules the APC measurements and communicates the measurement results to the data processing unit. The data processing unit converts the characteristics of the measurement data into a recipe comprising instructions to the lithographic apparatus. This method is very effective in suppressing the drift phenomenon associated with the lithographic apparatus.

Processing metrology data to produce correction operations performed by processing devices is important in semiconductor manufacturing. In addition to metrology data, characteristics and other contextual data of each patterning device, substrate, processing apparatus may also be needed to further optimize the manufacturing process. A framework in which available metrology and context data is used to optimize the lithography process as a whole is often referred to as a part of holistic lithography. For example, contextual data related to CD errors on a reticle can be used to control various apparatus (lithographic apparatus, etch stations) such that the CD errors do not affect the yield of the manufacturing process. Subsequent metrology data can then be used to verify the effectiveness of this control strategy, and further corrective actions can be determined.

Automated process control is often aimed at controlling, eg, reducing, lot-to-lot variations in substrate fingerprints in process parameters such as overlay, focus, and CD. A "fingerprint" is the variation (or error in a parameter) of a parameter over a region. In-field fingerprints are variations across fields, and in some cases will be the same for all fields on the substrate. The substrate fingerprint is the variation across the entire substrate. In some cases, the substrate fingerprint may be separated into an in-field fingerprint and an inter-field fingerprint. The present invention relates to all types of fingerprints.

As shown in Figure 2, conventional APC correction in high volume manufacturing (HVM) conditions is based on feedback. Measurements obtained from the plurality of preceding substrates W _N-1 through W _Nx are used to control one or more process parameters of the current substrate W _N . Measurements, or correction parameters derived from individual measurements, are grouped together and fed to a moving average (MA), for example an exponentially weighted moving average (EWMA).

In more complex cases, specific MAs are applied for inter-field and intra-field corrections, and other types of MAs are applied for higher-order corrections (eg, per-field corrections). In a more complex case, if two layers are expected to exhibit very similar fingerprints, one layer is fed with some of the corrections determined for the previous layer. However, this scheme also has several disadvantages.

First, a limited time filtering algorithm is used. The inventors have determined that the time-variation of a parameter cannot be accurately captured using a moving average.

Second, unnecessary metrology steps are often performed. In some cases, higher-order fingerprints (eg, in-field fingerprints) change more slowly than inter-field fingerprints. For example, the stabilization period of the etcher (causing an inter-field fingerprint) and the stabilization period of the projection system of a lithographic apparatus (causing an in-field fingerprint) can be very different, and thus the time variation of the corresponding fingerprint (time variation difference) will be different. Therefore, because all lots are measured with the same sampling designed to correct all substrates and in-field fingerprints, time is wasted that could be used for other purposes.

Third, existing automated process control methods do not use information from substrate processing tools. The inventors have determined that the root cause of the temporal variation of certain modeled parameters may be tied to the process tool. For example, time variations in wafer scaling can be linked to the stabilization of an etcher whose characteristics are determined by time variations in its sensor data. In such cases, using readily available etcher tool sensor data is much easier and less expensive than measuring more lots and/or wafers with a metrology tool to fine filter the temporal filter.

Therefore, embodiments of the present invention can alleviate this disadvantage, fine-tuning the APC feedback loop for reduced lot-by-lot variations in process parameters and/or allowing for reduced metrology sampling rates and/or densities. means are provided. One embodiment is shown in FIG. 3 , which shows metrology measurements obtained from a plurality of preceding substrates W _N-1 to W _Nx , used in conjunction with state data 200 to control the current substrate W _N . The state data 200 relates to conditions present in the associated layers on the substrates W _N-1 through W _Nx as they are processed by one or more process tools, such as an etcher or annealer.

In one embodiment of the present invention, information derived from metrology measurements may be provided in the form of a data time series, ie a series of data values each associated with time. It should be noted that the time associated with any data value does not necessarily have to be the time at which the measurement was performed, rather it is the time at which the associated manufacturing step, eg, exposure, was performed on the measured structure or target. The purpose of providing metrology steps and metrology data as a time series is to infer time variations in conditions prevailing in a manufacturing tool, for example a lithographic apparatus. The information derived from the metrology measurements may be the actual measurement results themselves or modeled parameters derived from the actual measurement results, such as transformations, rotations, scaling, etc.

State data related to the prevailing conditions in the manufacturing tool may also be provided in time series for the same purpose. State data may include control values applied to the manufacturing tool or measures of conditions prevailing in the manufacturing tool. In the latter case, the time associated with the state data value may be the time at which the measurement was made.

In one embodiment of the present invention, the automated process control system independently applies a time filter to different controllable process parameters. In the simplest embodiment, the user can decide for each controllable process parameter whether to apply a filter to each data time series. A library of smoothing filters is provided from which the user can select, for example:

- Bessel filter

- Butterworth Filter

- Matched filter

- Elliptical filter (Cauer filter)

- Linkwitz-Riley filter

- Chebyshev filter

- Biquad filter

- high-pass filter

- Low-pass filter

- band-pass filter

- Infinite impulse response filter

- Finite Impulse Response Filter

- Bilinear transformation

- Kalman filter

- Savitzky-Golay filter

It is also possible to use multiple filters connected in series or parallel to filter the input for a single controllable process parameter. In embodiments, a first filter is applied to the measurement data of the measurement data time series related to a first region of a substrate, and a second filter different from the first filter is applied to the measurement data of the measurement data time series related to a second region of the substrate applies to For example, measurement data related to an edge die may be treated differently than measurement data related to an inner die.

Figure 4 illustrates another embodiment of the present invention in which a plurality of lots (A ... N ... X) are processed using the same recipe, each lot comprising a plurality of substrates. After the lithography step and one or more process steps, metrology measurements are performed on the substrates of one or more lots, eg, lots A through M. Metrology measurements from some or all of lots A through M are used to calculate corrections to be applied to subsequent lots N using a mathematical model comprising multiple terms, eg, a polynomial. The various terms of the polynomial are calculated from metrology measurements from lots A through M using respective time filters 210 . The polynomial may include terms that are powers of coordinates (eg, x, y) representing positions on the substrate. The temporal filter may be different for each term of the polynomial. Metrology Measurements are performed on substrates in lot N, corresponding to those performed on some or all of lots A-M. It is possible to refine the model as more information becomes available. The refined model is used to determine corrections for subsequent lots (such as X).

Fig. 5 shows another embodiment of the present invention, similar to the embodiment of Fig. 4, except as described below. In the embodiment of Figure 5, state data from one or more process steps is used to fine-tune the filter 210 applied to the metrology data.

Fig. 6 shows another embodiment of the present invention, similar to the embodiment of Fig. 4, except as described below. In the embodiment of Figure 6, state data from one or more process steps is primarily used to determine corrections to be applied to subsequent lots, and an appropriate temporal filter 220 is applied to the state data. Metrology measurements are used to verify and protect against excursions. Therefore, fewer metrology measurements need be performed compared to the case where metrology is used as the primary determinant of corrections to be applied to subsequent lots. The number of metrology measurements performed per substrate need not be constant. This saves time and thus improves throughput.

7 shows an embodiment where only the state data is filtered by filter 220 and used to determine corrections for subsequent lots. No metrology data is used, so the metrology step can be omitted. This approach is particularly useful in the ramp-up step prior to mass manufacturing, where the test substrate can be cycled through process steps.

8 shows a process according to another embodiment of the present invention for allowing a user to set up an appropriate filter for an APC loop. In S1, initial data is obtained from the resource cluster. The initial data may include metrology data and/or status data from one or more substrates of one or more lots. The initial data may be referred to as training data. In S2, the initial data is processed so that time variations of sub-processes of the overall manufacturing process can be determined. In one embodiment, this may be done by determining the power spectral density (PSD), or similar graph, using a Fourier transform or other similar transform. This transformation transforms the initial data from time series data to frequency space data. The frequency spatial data is used to find the optimal time filter to use for each process parameter for the APC modeled parameter data flow. In one embodiment, the optimal temporal filter may be determined by an algorithm. Alternatively, a software interface is provided for receiving user input for selecting a filter in step S3.

One approach to selecting an appropriate filter is to determine the correlation between the particular process tool parameter and the APC model parameter (S4). In order to accurately pair up the process tool parameter(s) with metrology-based parameter(s) based on a shared time dependence, the correlation is a spectrum of power or energy density, or correlation coefficient matrices or similar. can be determined using As shown in Figure 9, a correlation is observed between the temperature T (shown at the top) in a process tool, e.g. an annealing furnace, and an overlay fingerprint (shown at the bottom) measured by a metrology tool. What can be can be possible This correlation can be provided to the user, allowing the user to decide which process tools and which APC parameters to pair up for a shared time dependency.

In addition, an embodiment of the present invention may advise and/or order the optimal time filters for selection parameters for optimal feedback control, and advise which process tool parameters to pair up for further fine tuning. can

In addition, an embodiment of the present invention may propose a time and wafer layout scheme for metrology sampling. For example, it may be desirable to measure a larger number of substrates at fewer points for inter-field fingerprint correction only, and to measure fewer substrates or lots with a denser measurement scheme for in-field correction.

Accordingly, preferred features of the present invention are as follows:

- APC filter settings can be specifically determined for the process steps involved

- For fine-tuned control, it is possible to find and correlate relevant process tool sensor data to APC model parameters.

- This can be done independently for each APC parameter

- Provide instrumentation sampling advice based on the aforementioned process.

Examples of process parameters to which the present invention can be applied include overlay, CD, CDU, sidewall angle, line edge roughness and focus. Suitable markers and measurement techniques for measuring these parameters are known in the art.

Although specific embodiments of the invention have been described above, it will be understood that the invention may be practiced otherwise than as described.

One embodiment is a computer program comprising one or more sequences of machine-readable instructions configured to instruct various apparatus as shown in FIG. 1 to perform measurement and optimization steps as described above and to control subsequent exposure processes. may include Such a computer program may be executed, for example, within the control unit LACU of FIG. 1 or the supervisory control system SCS or a combination of both. A data storage medium (eg, semiconductor memory, magnetic or optical disk) containing such a computer program stored therein may also be provided.

Although the above has specifically mentioned the use of embodiments of the invention in the context of optical lithography, the invention may also be used in other applications, for example imprint lithography, limited to optical lithography where the context permits. It will be acknowledged that this is not the case. In imprint lithography, the topography of a patterning device defines a pattern created on a substrate. The topography of the patterning device may be pressed into a layer of resist applied to a substrate, on which the resist is cured by applying electromagnetic radiation, heat, pressure, or a combination thereof. The patterning device is moved out of the resist after it is cured, leaving a pattern therein.

As used herein, the terms "radiation" and "beam" refer to ultraviolet (UV) radiation (e.g., having a wavelength of about 365, 355, 248, 193, 157, or 126 nm) and extreme ultraviolet (EUV) radiation ( for example, with wavelengths in the range of 1-100 nm), and all types of electromagnetic radiation, including particle beams such as ion beams or electron beams. Implementations of scatterometers and other inspection devices can be fabricated at UV and EUV wavelengths using suitable sources, and the invention is in no way limited to systems using IR and visible light.

The term “lens” as used herein, if the context permits, may refer to any or combination of various types of optical components, including refractive, diffractive, reflective, magnetic, electromagnetic, and electrostatic optical components. . Reflective components have the potential to be used in devices that operate in the UV and/or EUV range.

The scope and scope of application of the present invention should not be limited by any of the above-described exemplary embodiments, but should be defined only in accordance with the following claims and their equivalents.

Claims (15)

A device manufacturing method comprising:
acquiring a time series of measurement data of a plurality of substrates on which the exposure step and the process step have been performed;
obtaining a time series of state data related to conditions present in the process tool when the process step is performed on at least a portion of the plurality of substrates by the process tool;
applying a filter to the measurement data time series and the state data time series to obtain filtered data; and
using the filtered data to determine a correction to be applied to a subsequent exposure step performed on a subsequent substrate. The method of claim 1,
The method is
applying an additional filter to one or both of the measurement data time series and the state data time series to obtain further filtered data; and
using the further filtered data to determine further corrections to be applied to the subsequent exposure step performed on the subsequent substrate. 3. The method of claim 1 or 2,
Applying one or both of filters and additional filters is:
A method of manufacturing a device comprising applying a polynomial filter having a term in spatial coordinates of a substrate. 3. The method of claim 1 or 2,
Applying one or both of filters and additional filters is:
converting the measurement data time series and the state data time series into frequency space data;
applying a frequency filter to the frequency spatial data to obtain filtered frequency spatial data, and
converting the filtered frequency data into the filtered data. 3. The method of claim 1 or 2,
Applying one or both of filters and additional filters is:
applying a first filter to the measurement data of the measurement data time series related to the first region of the substrate; and
and applying a second filter different from the first filter to the measurement data of the measurement data time series related to a second region of the substrate. 6. The method of claim 5,
wherein the first zone is an edge zone and the second zone is an inner zone. A device manufacturing method comprising:
acquiring a time series of measurement data of a plurality of substrates on which the exposure step and the process step have been performed;
applying a filter to the measurement data time series to obtain filtered data;
using the filtered data to determine a correction to be applied to a subsequent exposure step performed on a subsequent substrate;
applying an additional filter to the measurement data time series to obtain additional filtered data; and
using the further filtered data to determine further corrections to be applied to the subsequent exposure step performed on the subsequent substrate. 8. The method of claim 7,
The method is
obtaining a time series of state data related to conditions present in the process tool when the process step is performed on at least a portion of the plurality of substrates by the process tool;
and applying a filter comprises applying a filter to the measurement data time series and the state data time series to obtain the filtered data. A device manufacturing method comprising:
acquiring a time series of measurement data of a plurality of substrates on which the exposure step and the process step have been performed;
obtaining a time series of state data related to conditions present in the process tool when the process step is performed on at least a portion of the plurality of substrates by the process tool;
converting the measurement data time series and the state data time series into frequency space data;
determining a filter to be applied to at least one of the measurement data time series and the state data time series based on the frequency spatial data to obtain filtered data;
applying the filter to the at least one of the measurement data time series and the state data time series to obtain filtered data; and
and using the filtered data to determine a correction to be applied to a subsequent exposure step performed on a subsequent substrate. 10. The method of claim 9,
the state data time series includes data related to a plurality of process steps performed on the substrate;
wherein determining the filter to be applied comprises selecting a subset of the state data time series related to the subset of process steps to be filtered and used to determine the correction. A device manufacturing method comprising:
acquiring a time series of measurement data of a plurality of substrates on which the exposure step and the process step have been performed;
obtaining a time series of state data related to conditions present in the process tool when the process step is performed on at least a portion of the plurality of substrates by the process tool;
converting the measurement data time series and the state data time series into frequency space data; and
determining, based on the frequency spatial data, a sampling scheme to be applied to a subsequent substrate to generate a measurement data time series. The method of claim 1,
wherein the correction is applied to correct at least one of overlay, dose and focus. 8. The method of claim 7,
wherein one or both of the correction and the further correction are applied to correct at least one of overlay, dose and focus. The method of claim 1,
wherein the process step is at least one process selected from the group of etching, annealing, implantation, deposition, and polishing. A computer readable recording medium storing a computer program comprising computer readable code means for instructing one or more lithographic tools to perform a method according to claim 1 .

Images